A highlight from Versatus - The Most Versatile DevEx in Web3
Hi everyone, Andy Pickering here, I'm your host and welcome to the Crypto Conversation, a Brave New Coin podcast where we talk to the people building the future in the Bitcoin, blockchain and cryptocurrency space. Hey team, we have a new sponsor here at the Crypto Conversation, BitGet, one of the world's leading copy trading cryptocurrency exchanges, yes indeed. What happens if you've got the funds to invest but you don't have the time to keep track of the market? You still want to make smart money moves, what do you do? Well copy trading is a popular choice for beginner traders. You can shorten your learning curve by uncovering tips and strategies from more experienced traders. BitGet's copy trading platform has over 80 ,000 elite traders to choose from and 380 ,000 followers just like yourself who are already using the BitGet copy trading platform as a potential passive income stream. All it takes is one click, you can subscribe to an elite profitable strategist, set your limits, automate your orders and monitor their trades. I've got some links in the show notes below, one link will take you through to the BitGet sign up page, give you a VIP discount. So learn all about it for yourself, thanks to BitGet. And now it is on with the show. My guest today is Andrew Smith, Andrew is the founder of Versatus Labs, building out the most versatile DevEx in Web3. Welcome to the show Andrew. Thanks for having me Andy. It is a pleasure, let's do what we do at the beginning of the show Andrew, it would be great if you could please introduce yourself. I'd love to hear a little bit about your, I guess, personal and professional backstory, what you've been doing that has led you to founding Versatus Labs. Yeah, absolutely. So I was born and raised in Miami, Florida, which is where I now reside again. I did do a stint in Denver, Colorado and an extended stint in Los Angeles. So I was gone from my hometown for about 12 years. I programming started at the age of 14, a technology teacher and seventh grade enemy, the classic, the C programming language book and said, learn this, I think it's going to be important. And so I did, never really did much as a kid other than like, you build like space invader clones and C and a couple of other things. Picked up Python and C++ a little bit later in life, during high school and, you know, was very, very interested in the cross -section of like machine learning and AI and economics. Economics is really sort of my first love, even though I'm a programmer, I kind of always wanted to be an economist, but just found that there's not really a lot of money in it unless you work for a political campaign. So it wasn't going to do that. And programming and machine learning in particular was something that I thought I could apply my love and knowledge of economics to. So it was building machine learning algorithms very, very early on before you add any of the sort of open source tools that you have today that makes it easy. And was sending my resume and GitHub around to a bunch of different hedge funds. Yes, this was going back about 10, 11 years now. And finally found one that was willing to give me a little bit of money to play around with. It's a group called Trident Asset Management. They're based part -time out of Connecticut and part -time out of Colorado, wasn't going to move to Connecticut. So that's what took me to Denver, then did the same thing for a fixed income shop based out of Newport Beach. That's how I ended up in Los Angeles. Started my first startup there, it's called Owl ESG, it's a environmental, social and governance data company built out, you know, some machine learning models and, you know, from PDFs, sort of scraping about 30 ,000 documents a day and extracting the data and building out a ESG data set. Grew that company and then in 2020 decided to start Versatus. So started this sort of hobby project, was doing a solo build on it, spent about 18 months solo building and was talking to a few friends in the space and they thought I was really onto something. So made some introductions, next thing you knew we were raising our first round from jumping big brain, hiring out an engineering team and now 14 months later, here we are. Very nice, very nice. Thank you, Andrew. Give us an idea then of, I guess, your vision for Versatus. What are you guys building? What's the vision? Yeah, so the vision is like the best way to put it, even though this is an imperfect if analogy is you think of like the cloud compute providers, AWS, Google Cloud, Azure, et cetera, you know, they own these huge data centers and these data centers are effectively a commodity business. You know, they build out a warehouse and put a bunch of servers in it, connect those servers to the Internet, occasionally maintain them and update them in and of themselves. They're not really that valuable. What makes them really valuable is that they provide all these tools that make it easy and efficient for developers to interact with those data centers and build applications on top of them to store data inside of them, et cetera. We believe that blockchain is analogous to that. It's not, again, it's an imperfect analogy. But if you kind of view the blockchains that exist in the world today and the ones that will come in the future as those data centers, next generation data centers where we provide value is we provide that program ability layer and compute layer that makes it easy and efficient for developers to build on top of blockchains. So we what we're building is a decentralized compute stack that enables developers to build in any language on any chain. And I think this is really powerful for a number of different reasons, which I'm sure we'll get to. But one of the major barriers to entry for developers is the language barrier. There's also a pretty big tooling barrier as well, which we saw that the language barrier, you know, if you're you want to build in Web3, the first thing you need to do is either go learn Solidity or Rust or one of the other languages. And Rust is a general purpose language. There are some people that already know it, but anybody that's entering into Web3 at the very beginning and they've got to go learn Solidity. Right. So a lot of them just don't view it as worthwhile to go learn Solidity. It's a domain specific language. The only thing you'll ever be able to do with that is build EVM compatible smart contracts. So until and unless there's a robust enough financial incentive for them to actually go and learn Solidity, they're probably not going to. But what we found from doing some pretty significant market research is if they could just use their existing languages and existing tools, they'd be happy to hobby hack and maybe even look for a job or start their own project and build on top of blockchains. So we want to make that process easier. We want to reduce the barrier to entry for developers. We believe that developers precede users, that you need developers to build applications that users actually want to use if we're ever going to see mass adoption for Web3. Yeah, I mean, that's a great point, Andrew. And I've seen you guys talk about this and some of your comms, I guess, because that's kind of it is flipping the script, right? Because everyone thinks, yeah, OK, it's the transition to Web3, easy as just got to build some user user friendly apps and and and if you build it, they will come. But of course, real life has has not been that simple. So so your philosophy is essentially the reverse of that. So you want to attract as many developers as possible. So just talk us through that again. I mean, you have a little bit, but just explain why you think that is really the key to the paradigm shift for Web2 to Web3. Yeah, absolutely. I think like just kind of telling the story of some case studies probably helps here, right? So you never know where a killer app is going to come from. I mean, Facebook started as a dating app for Ivy Leaguers, right? And it's Harvard and Yale dating app. You know, Slack started as a video game studio and Slack was their internal messaging network. So and now that is the product. Killer applications oftentimes come from experimentation. And the more experiments you have going on, the higher the probability that you're going to find stuff that people actually want to interact with and use. There are some precursors to what makes a killer app, things that make people's lives more convenient. That's just undeniably is going to make their life better, makes their work more productive. These are usually more business applications, makes the world more connected. These are social media type of applications or makes their life more affordable. So things that create efficiencies that reduce the cost of things that they were already doing. So, look, if I knew what that killer app was going to be, I'd probably go build that. It probably would be easier. But what I what I think where I think killer apps come from is lots of developers trying lots of things and competing for the limited funding and resources out there. And then you have unfortunately you do have gatekeepers in the world that you have VCs and you have investors and angel investors. So typically, yeah, there's going to be some stuff that's lost in the process of gathering funding and everything else that might have been really cool. But really, like if you have lots of things competing, probably the cream rises to the top and you're going to get well -funded, really interesting application ideas that can then promote themselves and attract users. The users are going to come for the applications right now. We have sort of the most users will ever have. If this is all we ever have to offer, which is effectively gambling and speculation, I think we've captured the gambling market pretty, pretty, pretty well. The speculator market we captured pretty well. They're here to make money off of token price fluctuations. If we want people that are here for the long term to use applications, well, we need to offer the applications that they want to use. And I think where that comes from, it's largely a numbers game. It's Pareto principle, you know, 10 percent, 20 percent of the developers are going to create the applications that get 80 percent, 90 percent of the users. So if we want to have a bigger 20 percent of applications that get lots of users, we need a bigger 100 percent. We need a bigger pie in general. And the only way to get a bigger pie is to reduce the cost, both time and money cost of building in Web3. And that's what we're attempting to do, particularly on the on the time cost of things, reduce the opportunity cost of learning how to build in Web3 by making it easier for them to build in Web3. So that's really sort of how we think about this. We think that developers necessarily are a precursor to users. If you look at like some of the market research we've done, it's kind of an either or like if there were more users, developers would take the time to learn this stuff. But the problem is, is that there's not going to be more users until developers learn how to build this stuff. So that's kind of where we see ourselves. We we believe we can be the catalyst for a Cambrian explosion of Web3 developers coming from all different walks of life, bring in product managers that they can understand how to manage a project that's being built in Python or Go or C++, but may not understand how to manage a project that's being built in Solidity, bring in on, you know, entrepreneurs that they come into this space and they look at, OK, well, how do I build a team out to build this? And what they see is extremely high cost of talent acquisition because there just isn't that big of a pool of Solidity developers. So make the talent pools that they can hire from significantly bigger, reduce that cost. Now you get some of those non -technical entrepreneurs looking at Web3 as a way to build their application. That's kind of the way we look at it. Just make the process easier, reduce those barriers. You'll get that first wave who's like jumping at the bit to come into Web3 and then they'll build some apps. You'll get more users. You'll then get the next wave of developers who see that there's financial incentives to doing so. It's going to be a process. It's going to take time. But we believe within the next seven to 10 years, if you offer up the correct tools and stacks, that about a third of all applications will be built on decentralized stacks for a number of different reasons, which we could talk to if you'd like. But that's where we see our value proposition is we make it easier for them. They come in, they build, then you get the users, then more come in and build, and so forth and so forth. You create a flywheel effect. OK, well, thank you, Andrew. And look, we don't need to get too deep into the weeds, but just talking about that decentralized stack, I suppose that you guys are building at Versatus. You have your own layer one blockchain, right? And there's the consensus mechanism, I believe, is proof of claim. So maybe just give us the kind of the two minute overview of your stack, I suppose. Yeah, so our L1 is primarily used for content addressing programs that are deployed to our network. So this is a way that our compute nodes can verify that they're executing the correct programs and such that watcher nodes and validators can also ensure that those compute nodes are not acting maliciously, that they're executing the correct programs. Our consensus mechanism, so proof of claim is actually our election mechanism. So this is how we elect nodes to quorums. Our consensus mechanism, we call it farmer harvester. Basically, it's a modification of what many distributed systems engineers would know as the worker collector model, but to fit a Byzantine fault tolerant model. So in your worker collector model, you basically have worker nodes that are individual nodes that they're allocated compute tasks. They execute those compute tasks and return the results to a collector node, which collects them and does batch updates into a database or to wherever they're storing state in our model. You don't want to have single nodes doing this work because then if a single node is malicious, they can actually create have state altering transactions that are incorrect. So we do have we form quorums as opposed to having single nodes. And then 60 percent of that quorum needs to what we call redundant, redundantly execute the program. So redundantly execute the program, return results, agree on results and then send votes to the what we call the harvester quorum. So, again, instead of having a single collector, we have a quorum of collectors that they then need to agree on the threshold of votes being reached before they would commit that to a block. So that's sort of very high level overview of how our architecture works. Now, again, like our goal is to enable language agnosticism on top of every chain. So not just for our L1, but on top of Ethereum, on top of other chains as well. And the primary reason for having our own L1 is it's a place where we can efficiently prove that compute nodes in our network are using the correct program, they're executing the correct program. And it's also a place where we can accrue value to those compute nodes. So whether they're being paid by another network's native token or they're being paid for executing compute on our network, we can emit our native tokens to them as an L1. So they're bootstrapped. And that way they're earning some money off of it. And then also it's a place where we can accrue fees back to our own L1 so that those compute nodes have a place where they're getting paid. Got it. Thank you, Andrew. If we kind of zoom out then to some more kind of, I guess, just a general state of where we are and the slow transition from Web 2 to Web 3. You saw a lot of the big brands, big financial institutions start to experiment with blockchain, but they were kind of like, they weren't really interested in building on Bitcoin or Ethereum. They went down the route of building their own private blockchains, which was a little bit pointless perhaps in hindsight. And now we're seeing with so many different chains around now and much more interoperability, brands and institutions are recognizing that it's to their benefit and everyone to build on the decentralized stacks that you're talking about. So maybe just you look at, I'd love you to paint a picture of, I suppose, your ideas of where we are now and your vision for what the next steps are just over, I guess, the next wave of adoption, maybe what's going to ignite the next hype cycle. How do you think about this? Yeah, so it's an interesting question. I try to steer away from predictions as much as possible. If I were a better investor, I probably would just be investing and making money that way. I do think the key, going back to hate to just sort of beat a dead horse, but the key is going to be getting more developers and whether those are enterprise developers, which I think what we're building provides a lot of value to enterprises. Again, they don't need to go out and hire a bunch of solidity developers that have four or five, six years experience. They can hire much more experienced developers or use the existing developers they have on staff. That to me is the key. I think we need more people trying things, pushing the limits of what's possible on top of this technology in order for us to find the use cases that are going to lead to mass adoption. I also think that enterprises, there are potentially some use cases for enterprise blockchains, but for the most part, I think one of the things that steered enterprises away from using public blockchains were privacy concerns. Right now, if you were to have a corporate wallet on top of Ethereum, everybody knows how much money you have in that. I think that level of transparency is something that scares a lot of enterprises and the closer we move towards being able to have on -chain privacy, so provability, but without revealing the underlying values, the more you'll see enterprises adopt public blockchains as a place, as a development environment, as a place to build and deploy applications to both internal applications as well as consumer facing or other business facing applications. But I think you've got to solve that privacy issue. Transparency is good when needed. It's also something that can be a deterrent to particularly large publicly traded companies who have to report to the SEC, who get audited, all these other things. They don't want all of this information, their financial information public. So finding ways to create some privacy around that I think will probably help with enterprise adoption. Yeah, yeah. Makes perfect sense, Andrew. What about, how does AI fit into this? I know it's a little bit of a tangent, but I've seen you guys talk a little bit about AI. I think you've probably got some opinions. So yeah, I mean, anything you want to kind of speculate on in terms of the, I guess the intersection of AI and web3 in the future? So in one word, trust, I think that's the key is that we're able to offer trust is very, very expensive. And I'm not talking about just necessarily blockchain trust, but trust in general. It's very expensive and it's at the core of how and why society works. If you don't have trust, society breaks down. So we have to trust each other, that we have our individual best interests in mind. And as a result of us trusting that we each want to do what's best for ourselves, we know that we're not going to put ourselves in a situation to damage each other because that might hurt ourselves. So having trust in AI models is going to be really, really important. And right now that mechanism works because OpenAI runs it and OpenAI is a big company, they have profit motives, but it's all centralized. As we move to a world where there's decentralized AI models, there needs to be some way to trust that that AI model is not malicious. And I think blockchain can be a huge component of that and tokenization, staking, and being able to lend trust to compute models is a really important component of it. I think it's an area where we fit in really, really well in particular. So that to me is the most obvious intersection of AI and blockchain. Particularly when it comes to things like deep fakes, I think you want to be able to have some verifiability behind images. You want to have some verifiability behind videos. You can just imagine a scenario where somebody creates a deep fake there's and no way to prove that this came from an AI model, and all of a sudden chaos ensues in a city or in a region or in a country because of some deep fake that people think is real. So there are a lot of concerns around fake news use cases for AI, and how do we solve for that problem? How do we put a marker on that image or on that video that proves that this came from a model and having some sort of watermark of trust? I think that crypto can provide that in some ways. So that's one area. I also think there's a lot of concern about existential threats related to AI and decentralizing AI models and getting them out of the hands of individuals and into the hands of communities, open sourcing them, and then providing incentives around building these models in a way to where they won't create existential threats. I don't think we're quite there yet. I'm less of an AI doomer than a lot of people. But to the AI doomers, I would say use crypto as a way to provide some of these guarantees that your model is not going to go off the rails.